16S Microbiome Assessment in Pneumonic Calves from a farm in BC Canada

Description

The primary goal of the study was to determine differences between the microbiomes of healthy (HLT), pneumonic (POS), and previously pneumonic (PRV) calves from a single dairy farm dealing with a high prevalence of Mycoplasmopsis bovis. 16S rRNA sequencing was completed by Norgen Biotek Corp. on the Illumina MiSeq platform. All fastq sequences generated are accessible on NCBI's Sequence Read Archive repository under the BioProject accession number PRJNA1389668. Amplicon sequence variants (ASVs) were constructed using DADA2 in QIIME2. Taxonomic classification was completed using the SILVA v138 reference database. Subsequent analyses were completed in R Statistics. This dataset includes the sample metadata (Supplementary Table 1), relative abundance distribution across genera and species (Supplementary Table 2), the differential abundance results from the ANCOM-BC2 analysis using the ANCOMBC v1.2.8 R package (Supplementary Table 3), and the taxa-taxa correlation results from the SECOM analysis which was also implemented in the ANCOMBC R package (Supplementary Table 4). Our study focused on Mycoplasmopsis bovis (M. bovis) abundance in pneumonic calves. Briefly, we found that M. bovis was more associated with the loss of commensal taxa that were enriched in healthy animals and was significantly enriched in pneumonia positive animals. Overall, these findings suggest that M. bovis is associated with dysbiosis within the respiratory microbiota and may influence BRD pathogenesis. ANCOM-BC2 data displays the log2 fold change, p-value, and BH-adjusted q-value as well as other differential abundance metrics for each comparison. For clinical group comparisons, healthy animals were the reference. SECOM data outlines the taxa-taxa abundance correlations for each taxa among the dataset. Both nonlinear (distance) and linear (Pearson) correlation was calculated and the associated BH-adjusted p-value.

Steps to reproduce

Raw read quality was assessed using FastQC (Andrews, 2010) and MultiQC (Ewels et al., 2016). Raw reads were processed to remove adapters and filter low-quality reads using the BBtools suite (Bushnell et al., 2017). Reads passing filters proceeded to targeted amplicon 16S metagenomics analysis using QIIME2 version 2024.10 (Bolyen et al., 2019). Samples were denoised and clustered into amplicon sequence variants (ASVs) using DADA2 (Callahan et al., 2016). Taxonomic classification of ASVs was performed using the SILVA v138 reference database (Quast et al., 2013) with a Scikit-learn-based Naïve Bayes classifier under a 70% confidence threshold (Pedregosa et al., 2010). QIIME2 was then used to export the resulting feature table and taxonomy table, which were imported into R v4.4.2 (R Core Team, 2024) with dplyr v1.1.4 (Wickham et al., 2023) to construct a phyloseq object (v1.50.0; McMurdie and Holmes, 2013). The presence of M. bovis DNA within samples was determined by both the uvrC-V2 qPCR assay and 16S rRNA sequencing ASVs classified as M. bovis. Fisher Exact Test was used to compare whether the prevalence of M. bovis was associated with pneumonia. In addition, differences in M. bovis abundance across clinical groups were evaluated using the Kruskal-Wallis test for both 16S-derived M. bovis abundances and the uvrC-V2 qPCR results. Differential abundance was assessed using Analysis of Compositions of Microbiomes with Bias Correction 2 (ANCOM-BC2) via the ANCOMBC v2.8.1 package (Lin and Peddada, 2020, 2024; Lin et al., 2022). The analysis was done at the species level, with filtering applied to exclude features present in fewer than 10% of samples and samples with fewer than 1,000 reads. The multivariate statistical model included both pneumonia status and body site as fixed effects, with clinically healthy samples as the reference group. Additionally the model controlled for repeated measures. ANCOM-BC2 calculates pairwise fold change between log-adjusted abundances. Log2 fold change (FC), p-values, and p-values adjusted for false discovery rate (q-value) using the Benjamini-Hochberg procedure were calculated for each taxon that had a statistically significant p-value in at least one of the comparisons (α = 0.05). Log2 fold changes of each taxon were visualized using a heatmap plotted using ggplot2 (Wickham, 2016). Taxa co-occurrence and abundance correlation patterns were evaluated using the Sparse Estimation of Correlations among Microbiomes (SECOM) method, also implemented in the ANCOMBC package. SECOM was applied using both distance correlation and Pearson correlation methods. Features were retained if present in a minimum of 10% of samples. Pseudo-counts were applied before transformations for zero handling and p-values were generated using 1,000 permutations and adjusted using the Benjamini-Hochberg procedure. Significant taxa associations (α = 0.05) were retained based on having a correlation coefficient greater than 0.2.

Institutions

• Kwantlen Polytechnic University

  British Columbia, Surrey

Categories

Licence